Rizel Scarlett is the principal developer advocate at Entire, building human-agent collaboration into the infrastructure of software development. Previously a software engineer, GitHub Copilot advocate, and technical lead for open-source developer relations at Block, she focuses on making AI-generated work understandable, reviewable, and accountable to human developers.
Scarlett worked service jobs and trained in phlebotomy before meeting a Black woman software engineer who introduced her to a profession she had barely known existed. While working at Massachusetts General Hospital, she studied computer information systems at Bunker Hill Community College and interned at the Federal Home Loan Bank of Boston. Hack.Diversity led to internships at HubSpot, and Resilient Coders provided full-stack development training.
She subsequently interned at Formlabs, worked as a software engineer at Veson Nautical and Hi Marley, and completed a computer-science degree at Boston University. Her account of entering software connects financial stability, representation, mentorship, and access to technical communities. She also volunteered, held programming leadership responsibilities, and later served as an adviser at G{Code} House, supporting women and nonbinary people of color entering technology.
At GitHub, Scarlett taught developers to use Copilot and wrote about practical developer tools, including Codespaces, browser extensions, automation, and open-source participation. She then joined TBD, a Block business unit, as a staff developer advocate working on decentralized identity, verifiable credentials, and personal data ownership. Her subsequent work at Block centered on the Goose open-source agent, configurable models, external tools, and inspectable agent workflows.
At Entire, she helped introduce an agent-oriented development platform that connects Git history with prompts, inspected files, tool activity, and other context behind AI-generated code. That record supports code review, agent handoffs, and accountability when conventional diffs cannot explain why a change was made.
- Thoughtful anthropomorphism: Calling an AI agent a software engineer inflates expectations and encourages employers to equate automation with human replacements. Scarlett favors precise product language that acknowledges missing context, hallucinations, and unreliable outputs while preserving developer trust, as she argues in her AI Engineer Summit talk.
- Developer control and transparency: Useful AI products expose limitations, offer configurable models, provide clear documentation, and make workflows inspectable. Scarlett expects developer advocates to become intensive users of their own tools, understanding context management, tool calling, and agent failure modes firsthand.
- Agent context as engineering infrastructure: Prompts, tool calls, inspected files, and session history give engineers information that code diffs omit, making agent-produced changes easier to review and trace.
- Multi-model implementation and review: Scarlett built a Goose Ralph loop that resets context between iterations and assigns implementation and review to separate models, creating explicit checks within an autonomous workflow.
